4.2.5 and 5.0.0

classic Classic list List threaded Threaded
13 messages Options
Reply | Threaded
Open this post in threaded view
|

4.2.5 and 5.0.0

Paul Franklin-5
I no longer think 5.0.0 should be released "soon".

Since some developers still think there are too many
problems in the DB-API code in master, I think it makes
much more sense to follow Nick's original plan (as he
said in the April message I append below) and release
5.0.0 in 2017.

But there have been many features added to master
which all of us who have added them would like to see
used, features which have nothing to do with the DB.

So I instead propose that master should be copied to
a new gramps43 branch instantly, then the DB-API code
removed from it, not just disabled as now, to differetiate
it from master.  Then a new gramps can be released
essentially immediately --4.3.0 -- after only a short
"string freeze" interval.

Which is what Nick said last April that we should do.


---------- Forwarded message ----------
From: Nick Hall <[hidden email]>
Date: Sat, 16 Apr 2016 15:25:25 +0100
Subject: Re: [Gramps-devel] Preparation for Gramps 5.0
To: [hidden email]

Paul,

A sensible alternative would be to release v4.3 in June, with BSDDB as
the default database.

The Select API / QuerySet interface should also be removed.  As I have
already mentioned, agreement was never reached on this.

Efficient select functionality could then be fully integrated into the
filters for a release of v5.0 in June 2017.


Nick.

------------------------------------------------------------------------------
_______________________________________________
Gramps-devel mailing list
[hidden email]
https://lists.sourceforge.net/lists/listinfo/gramps-devel
Reply | Threaded
Open this post in threaded view
|

Re: 4.2.5 and 5.0.0

DS Blank
On Tue, Sep 13, 2016 at 3:59 PM, Paul Franklin <[hidden email]> wrote:
I no longer think 5.0.0 should be released "soon".

Since some developers still think there are too many
problems in the DB-API code in master, I think it makes
much more sense to follow Nick's original plan (as he
said in the April message I append below) and release
5.0.0 in 2017.

As I mentioned, there are no blockers in the roadmap, and not even anything related to DBI.  What are these mysterious problems to hold up a release? 
 

But there have been many features added to master
which all of us who have added them would like to see
used, features which have nothing to do with the DB.

So I instead propose that master should be copied to
a new gramps43 branch instantly, then the DB-API code
removed from it, not just disabled as now, to differetiate
it from master.  Then a new gramps can be released
essentially immediately --4.3.0 -- after only a short
"string freeze" interval.

That sounds like an incredibly bad idea... one doesn't want a ton of such changes in a point release. This is exactly why we have integer releases. And, you would make moot all of the testing we did in the alpha.
 
-Doug

------------------------------------------------------------------------------

_______________________________________________
Gramps-devel mailing list
[hidden email]
https://lists.sourceforge.net/lists/listinfo/gramps-devel
Reply | Threaded
Open this post in threaded view
|

Re: 4.2.5 and 5.0.0

Paul Franklin-5
On 9/13/16, Doug Blank <[hidden email]> wrote:
> As I mentioned, there are no blockers in the roadmap, and not even anything
> related to DBI.  What are these mysterious problems to hold up a release?

I see lots of things on the roadmaps, for both "5.0.0" and
"5.0.0-alpha2" and even "4.2.5"  -- so what roadmaps are
you looking at?  Furthermore, I see six bugs clearly stated
to be about the DB layer, including one marked DB-API.

> That sounds like an incredibly bad idea... one doesn't want a ton of such
> changes in a point release. This is exactly why we have integer releases.
> And, you would make moot all of the testing we did in the alpha.

A "4.3.0" release would be just as much of a a major
release as "4.2.0" or "4.1.0"  so I don't know what you
are talking about.

And your statement about the 5.0.0-alpha1 code is
likewise totally wrong, since all the code which was
tested in that release will be in the 4.3.0 release I
proposed -- except the DB-API code.

------------------------------------------------------------------------------
_______________________________________________
Gramps-devel mailing list
[hidden email]
https://lists.sourceforge.net/lists/listinfo/gramps-devel
Reply | Threaded
Open this post in threaded view
|

Re: 4.2.5 and 5.0.0

Nick Hall
In reply to this post by Paul Franklin-5
On 13/09/16 20:59, Paul Franklin wrote:
> I no longer think 5.0.0 should be released "soon".

I share your concerns.  I don't think that the DB-API code is of a high
enough standard for release yet.


>
> Since some developers still think there are too many
> problems in the DB-API code in master, I think it makes
> much more sense to follow Nick's original plan (as he
> said in the April message I append below) and release
> 5.0.0 in 2017.

Yes.  A bad release would damage the reputation of Gramps.


>
> But there have been many features added to master
> which all of us who have added them would like to see
> used, features which have nothing to do with the DB.
>
> So I instead propose that master should be copied to
> a new gramps43 branch instantly, then the DB-API code
> removed from it, not just disabled as now, to differetiate
> it from master.  Then a new gramps can be released
> essentially immediately --4.3.0 -- after only a short
> "string freeze" interval.
>
> Which is what Nick said last April that we should do.


I don't see a problem with releasing v4.3 with the DB-API code removed.  
Obviously, we would still need to carry out some thorough testing.

Are you able to do the necessary work?


Nick.



------------------------------------------------------------------------------
_______________________________________________
Gramps-devel mailing list
[hidden email]
https://lists.sourceforge.net/lists/listinfo/gramps-devel
Reply | Threaded
Open this post in threaded view
|

Re: 4.2.5 and 5.0.0

Paul Franklin-5
On 9/15/16, Nick Hall <[hidden email]> wrote:

> On 13/09/16 20:59, Paul Franklin wrote:
>> I no longer think 5.0.0 should be released "soon".
>
> I share your concerns.  I don't think that the DB-API code is of a high
> enough standard for release yet.
>
>
>>
>> Since some developers still think there are too many
>> problems in the DB-API code in master, I think it makes
>> much more sense to follow Nick's original plan (as he
>> said in the April message I append below) and release
>> 5.0.0 in 2017.
>
> Yes.  A bad release would damage the reputation of Gramps.
>
>
>>
>> But there have been many features added to master
>> which all of us who have added them would like to see
>> used, features which have nothing to do with the DB.
>>
>> So I instead propose that master should be copied to
>> a new gramps43 branch instantly, then the DB-API code
>> removed from it, not just disabled as now, to differetiate
>> it from master.  Then a new gramps can be released
>> essentially immediately --4.3.0 -- after only a short
>> "string freeze" interval.
>>
>> Which is what Nick said last April that we should do.
>
>
> I don't see a problem with releasing v4.3 with the DB-API code removed.
> Obviously, we would still need to carry out some thorough testing.
>
> Are you able to do the necessary work?
>
>
> Nick.
>
>
>

Yes, I am willing to make a new gramps43 branch.

As a matter of fact I have been working on that, as a
proof of concept, and I am mostly done.  At least with
the things I was able to guess were DB-API things.

But as I am not an expert on the DB layer I would like
to coordinate my work with somebody who knows more.

I was thinking of Tim since he filed the DB bugs on
the bug tracker.  I hope he will be able to help.

So one possibility would be to wait for Tim to be able
to help, before I make the gramps43 branch.  Another
possibility would be for me to create the gramps43
branch right now, and upload it to github, so anybody
could see it, try it, test it, modify it, etc.

I prefer the latter choice but that would mean that all
developers would have to commit any bug fixes to both
master and gramps43, once it existed, and while I am
of the opinion that is trivial, in the past others wanted
the new maintenance branch made at the last minute.

But for the record I am unable to do other related things,
for instance moving bugs on the roadmap "down the
line" to the next release -- or even moving fixed bugs to
a new gramps43 roadmap on the bug tracker.  Or doing
whatever notification to our translators usually is done
before we make a major release.

So I assume others will (continue to) do things like that.

------------------------------------------------------------------------------
_______________________________________________
Gramps-devel mailing list
[hidden email]
https://lists.sourceforge.net/lists/listinfo/gramps-devel
Reply | Threaded
Open this post in threaded view
|

Re: 4.2.5 and 5.0.0

manzi.sam
In reply to this post by Nick Hall
CONTENTS DELETED
The author has deleted this message.
Reply | Threaded
Open this post in threaded view
|

Re: 4.2.5 and 5.0.0

manzi.sam
CONTENTS DELETED
The author has deleted this message.
Reply | Threaded
Open this post in threaded view
|

Re: 4.2.5 and 5.0.0

Paul Franklin-5
In reply to this post by manzi.sam
On 9/15/16, Sam Manzi <[hidden email]> wrote:
> I vote to continue with the 5.0 release and believe it makes more sense to
> leave DB-API code in and only make it available when running in debug mode?

The /only/ difference between the proposed gramps43
branch (and thus 4.3.0) and the present master will be
the lack of DB-DPI code in gramps43.

So you are really saying you want the DB-API code -- even
though other developers are saying it is not ready yet?

It won't be removed from master.  As a developer, you know
how to run master, with DB-API turned on if you want it that
way, so I wonder why that isn't sufficient for you?

We are talking about the effect on users, not developers.

------------------------------------------------------------------------------
_______________________________________________
Gramps-devel mailing list
[hidden email]
https://lists.sourceforge.net/lists/listinfo/gramps-devel
Reply | Threaded
Open this post in threaded view
|

Re: 4.2.5 and 5.0.0

Paul Franklin-5
In reply to this post by manzi.sam
On 9/15/16, Sam Manzi <[hidden email]> wrote:
> The following would need to be hidden:
> * Preferences > Family Tree > Database backend
> * Manage Family Trees > Convert (button)  & possibly "Info: (button)
> * Manage Family Trees > Database Type (Column)

I don't see any reason to remove the FTM's Info button,
nor the Database Type column (it will just never change),
but I have already removed the other two, in my prototype.
But if others think the Info button and/or DB Type column
should be removed, I will not stand in their way.

------------------------------------------------------------------------------
_______________________________________________
Gramps-devel mailing list
[hidden email]
https://lists.sourceforge.net/lists/listinfo/gramps-devel
Reply | Threaded
Open this post in threaded view
|

Re: 4.2.5 and 5.0.0

Benny Malengier
In reply to this post by Nick Hall

> But there have been many features added to master
> which all of us who have added them would like to see
> used, features which have nothing to do with the DB.

 
Not having followed development, what do you consider the main features added apart from the DB changes?

The commit list is somewhat long to weed through, and the roadmap does not seem to list big features

Benny


------------------------------------------------------------------------------

_______________________________________________
Gramps-devel mailing list
[hidden email]
https://lists.sourceforge.net/lists/listinfo/gramps-devel
Reply | Threaded
Open this post in threaded view
|

Re: 4.2.5 and 5.0.0

Paul Franklin-5
On 9/15/16, Benny Malengier <[hidden email]> wrote:
> Not having followed development, what do you consider the main features
> added apart from the DB changes?


Hi Benny,

I haven't tried to work on the NEWS file for the gramps43
branch, since I assume it is made in some semi-automated way
and I don't know how to do that.  I'll leave that to Jerome,
when he starts getting ready to release 4.3.0.

So this is from memory and mostly from my own point of view,
although of course others have made lots of changes too, bug
fixes especially.  (These are not in any order, by the way.)

One of the big things which happened recently is that Serge
has added the translated-output option to the NarWeb report,
which people have wanted, and also to the WebCal report.
(He's fixed many problems in the NarWeb report also.)

I added the option to most reports to be able to filter the
output based on "living" people, like in the NarWeb report.

The book (report) was augmented so that "paper" options (as
well as the filename) are now saved in books.xml too.  And
more report "subjects" were added to the book (GUI) also.

For our Windows users, many of the dialogs were given a
transient parent, although doubtless more are needed.

Nick added a few "family" filters and so the Family Group
report was augmented to use them, for a feature request.  In
the process, people filters now sort on full names.

Lots of modules had their pylint scores improved, by several
people, although we all hope no user will notice that.  8-)

Likewise when Nick changed several classes to be "abstract".

Doug added a CacheDB layer to reports and improved the time
for a NarWeb run from two hours to eight minutes (IIRC).
(It will be in gramps43, so I mention it.)

Tim added a DummyDB class to hopefully prevent some DB
startup errors.  (It will be in gramps43, so I mention it.)

PaulC has added many fixes to the importers, especially the
GEDCOM one, as well as unittests for many things, but I am
not sure how many were also put into gramps42.

Serge added a tool to detect loops in family trees, but
again I am not sure if it was also put into gramps42.

I'm sure I've forgotten many things so don't assume this
list is a complete one, as I said earlier.

I will also offer the observation, which you probably know
better than I do, that for a few years now I don't think
there was any particular "big push" for each year's major
release.  I don't recall thinking that since your 4.0.0.  I
would describe the time since 4.2.0 as gradual evolution.
(If we exclude the DB-API work, which we will for 4.3.0.)

HTH.

------------------------------------------------------------------------------
_______________________________________________
Gramps-devel mailing list
[hidden email]
https://lists.sourceforge.net/lists/listinfo/gramps-devel
Reply | Threaded
Open this post in threaded view
|

Re: 4.2.5 and 5.0.0

Benny Malengier
Ok, nice overview.

Jerome extracts commit messages, but in the end must rework it to a real text for the release notes, so this list of you helps.

I see on the user list a person indicating to wait for 5.0 so as to use Gramps online version, so some users are aware and waiting for that as a feature.
On the eastman blog the first comment is also about gramps-connect not being officially announced, so again a person being aware of it, and hoping it to help in some way in the workflow.

Gramps-connect needs the DB-API. The previous approach has been abandoned because of too much overhead. Obviously, Gramps-connect is not Gramps proper, more a growing of the ecosystem, but I don't see a need to block it too, I now have a lot of chromebooks in the house for the children and Gramps is not working on it if I don't hack them.

All in all, I would not too quickly push for a 4.3 without the DB-API, but more take an approach as Sam advices, with DB-API present in the background already.
Looking at the roadmap, I don't see actual bugs with the DB-API at the moment. Issues Tim mentions is stuff which was already ugly code before DB changes, it relates to organically grown code to avoid issues on startup, empty databases and such. Things that can be tackled in point updates, they should not hold up a release.

It seems to me there is more need of a release manager that selects what still has to be done for 5.0, then does the work to get that, and move the rest to 5.1.
So on the roadmap, I would move all alpha2 issues to 5.0.0, create a 5.0.1 and 5.1.0, and remove the alpha2 as obsolete.
Then move things that will not be done to 5.1.0, and things which are bugs but not blocking a release to 5.0.1. Then see what still is needed and ask for developers to assign bugs to them to get them off the roadmap. A beta or alpha of 5.0.0 can still be done based on status of the 5.0.0 roadmap, but I see alpha's and beta's more as an intermediate state towards 5.0.0, not really something that needs an own roadmap. So I would advise Jerome to spin off an alpha2 or a beta of 5.0.0 next week. Alpha if we assume it will still eat your data, beta if we assume it will not destroy the data but still crash here and there.

I have too many other projects running at this time to jump in with programming, so I can only advise on my take on the situation.

Benny

2016-09-16 20:28 GMT+02:00 Paul Franklin <[hidden email]>:
On 9/15/16, Benny Malengier <[hidden email]> wrote:
> Not having followed development, what do you consider the main features
> added apart from the DB changes?


Hi Benny,

I haven't tried to work on the NEWS file for the gramps43
branch, since I assume it is made in some semi-automated way
and I don't know how to do that.  I'll leave that to Jerome,
when he starts getting ready to release 4.3.0.

So this is from memory and mostly from my own point of view,
although of course others have made lots of changes too, bug
fixes especially.  (These are not in any order, by the way.)

One of the big things which happened recently is that Serge
has added the translated-output option to the NarWeb report,
which people have wanted, and also to the WebCal report.
(He's fixed many problems in the NarWeb report also.)

I added the option to most reports to be able to filter the
output based on "living" people, like in the NarWeb report.

The book (report) was augmented so that "paper" options (as
well as the filename) are now saved in books.xml too.  And
more report "subjects" were added to the book (GUI) also.

For our Windows users, many of the dialogs were given a
transient parent, although doubtless more are needed.

Nick added a few "family" filters and so the Family Group
report was augmented to use them, for a feature request.  In
the process, people filters now sort on full names.

Lots of modules had their pylint scores improved, by several
people, although we all hope no user will notice that.  8-)

Likewise when Nick changed several classes to be "abstract".

Doug added a CacheDB layer to reports and improved the time
for a NarWeb run from two hours to eight minutes (IIRC).
(It will be in gramps43, so I mention it.)

Tim added a DummyDB class to hopefully prevent some DB
startup errors.  (It will be in gramps43, so I mention it.)

PaulC has added many fixes to the importers, especially the
GEDCOM one, as well as unittests for many things, but I am
not sure how many were also put into gramps42.

Serge added a tool to detect loops in family trees, but
again I am not sure if it was also put into gramps42.

I'm sure I've forgotten many things so don't assume this
list is a complete one, as I said earlier.

I will also offer the observation, which you probably know
better than I do, that for a few years now I don't think
there was any particular "big push" for each year's major
release.  I don't recall thinking that since your 4.0.0.  I
would describe the time since 4.2.0 as gradual evolution.
(If we exclude the DB-API work, which we will for 4.3.0.)

HTH.


------------------------------------------------------------------------------

_______________________________________________
Gramps-devel mailing list
[hidden email]
https://lists.sourceforge.net/lists/listinfo/gramps-devel
Reply | Threaded
Open this post in threaded view
|

Re: 4.2.5 and 5.0.0

jerome
In reply to this post by Paul Franklin-5
Yes, we can quickly see changes against master branch.

e.g.,
$ git checkout master
$ git log v4.2.4..

but to understand all messages for generating a right
list can be sometimes a little bit difficult!
Agreed, need to look at commit, which can be also difficult
to understand. And are they major or minor? Feature or for future?

Jérôme

--------------------------------------------
En date de : Ven 16.9.16, Benny Malengier <[hidden email]> a écrit :

 Objet: Re: [Gramps-devel] 4.2.5 and 5.0.0
 À: "Nick Hall" <[hidden email]>
 Cc: "Gramps Development List" <[hidden email]>
 Date: Vendredi 16 septembre 2016, 8h47
 
 
 > But there have been many features added to master
 
 > which all of us who have added them would like to see
 
 > used, features which have nothing to do with the DB.
 
  
 Not having followed development, what do you
 consider the main features added apart from the DB changes?
 
 The commit list is somewhat long to weed through,
 and the roadmap does not seem to list big features
 
 Benny
 
 
 -----La pièce jointe associée suit-----
 
 ------------------------------------------------------------------------------
 
 -----La pièce jointe associée suit-----
 
 _______________________________________________
 Gramps-devel mailing list
 [hidden email]
 https://lists.sourceforge.net/lists/listinfo/gramps-devel
 

------------------------------------------------------------------------------
_______________________________________________
Gramps-devel mailing list
[hidden email]
https://lists.sourceforge.net/lists/listinfo/gramps-devel